Support to build a device with multiple IMUs

[DUTCHnotDUTCH]

Hello All,
Apologies in advance for sounding like a noob (that is because I am). 
I am building a wearable device as a hobby, and I need a 6-axis gyroscope with IMU. Ideally, the device should have more than one IMU, but considering that I have absolutely no idea how to even assemble the device, I am reaching out to this forum to hopefully get an initial idea of what it is I need, so as to hopefully set me off in the right path. 

The sensors I am thinking in using: Initially I have been looking at ASM330LHHTR and A3G4250DTR, strictly because they are small. The idea I have is to begin testing with a simple setup connected to power, but to eventually make it battery-powered.

A bit more information about what I want to measure: eventually I want to set up a strip of these IMUs in a straight line, and have them measure at the same time, and feed it to something like a raspberry pi so as to transmit BT signal. 
To make a setup where I have IMU --> [something] --> BT-Signal-Generator, I can obviously argue back and forth with AI, but I was hoping to get more specific assistance from anyone who may have used these kinds of sensors in the past and who may know exactly what I need for that [something] part. I understand I may need a multiplexer if I am using multiple IMUs (or do I?), to integrate the different IMU signals and discriminate each sensor. 

Why am I even here asking for help: as I previously established, I have no idea how to do this, but I think I understand in theory what I need to do it. What I would like to know is the following:
1. Are there any extra/miscellaneous parts that I am not aware of, and that I should include in my setup to achieve my desired purpose? Specifically, whatever device/s should go into the [something] part (mentioned above) 
2. Considering the setup mentioned, is there a better/more appropriate IMU/gyroscope that is recommended? If so, how would this change the setup in (1)?

Thank you very much for y'all's time.

[Smithson]

It is called the Sensor Tile Box Pro 

[DUTCHnotDUTCH]

Thanks for your reply.
I have certainly seen that one, and it is something I am considering to perhaps start building the software, but the problem is that assembling a system with multiple of these is much more expensive than needed, but also includes a lot of sensors that I am not particularly interested in using — hence why I intended to build something from scratch.
Do you have experience with assembling the sensors individually into their EVAL boards?

[Smithson]

I am evaluating the MKI210V2k passing through the MKI109V3.

This is a good combination, but we are looking to monitor bridges, so our requirements are different.  

I am not someone who will assemble units, that is not my job.  

I look at the performance of units,

One big concern is the thermal layer,  which is however absolutely necessary for our analysis.  We just look at the whole maths.  The older accelerometers, circa 2000, have a thermal f to the power -n, which is usually about 1.4, so on the large bridges say across the Yangzte River first modes are buried in the low thermal zone,  not a problem if you know how to look.  

The MKI210 has a much flatter thermal layer, but it is a pain to keep running to a TRL 7 level, TRL 2 is easy, 7 is a challenge.  But it is a good device as long as you add a USB managed hub to it,   It took me three months to write the software for the 210.  The 210 is not stable in output frequency, once I found that out, it was a bit more work, but the results look ok, it is the results over years that matter. 

The Sensor Tile Box Pro is a real challenge, it uses WINUSB and the software is locked.  But it has a super flat thermal response.  I have not compared it to the 210 yet, but compared to the CX1 it is a dream.  The frequency is a bit more stable, but the on-off issue remains.  

So I am not much help.  

[Smithson]

I ask questions on this forum, more as an aid to myself, I really do not expect answers to the sort of stuff I do, 

so the Sensor Tile Box. 

The NPP is as expected, our team lives and dies in the purple region.  

This is a pure a thermal layer as one can get on human earth, without spending a lot of money.  The only thing you can tell from this thermal FFT data is the temperature,  LOL an Iowa engineer who works with me dug a hole in his back yard in the middle of a winter storm and we got the necessary data to ensure that was a correct statement.  He was not amused, I just had to chuckle.   The analysis of the linear regression using EXCEL, shows the line is "a straight line" and the t Stats say it is accurate and the P Values are a waste of time as a modern computer cannot represent the p values from the T Stat, trust me I have tried and failed there, but anything over 2 is ok and and 14 means it is as certain as we get on this earth.  

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%
Intercept	0.024229647	0.000179833	134.7342586	0	0.023877155	0.024582139
X Variable 1	-4.6482E-06	3.18852E-07	-14.57791477	7.73748E-48	-5.27318E-06	-4.02321E-06

And we take the 18 days of collected data from the sensor, 

For the X axis - 842.  And my work mate in England, she says garbage and I say wait, it will take time to see if the device is sound.  By the way -842 is not good for the structure but it is just a block under my desk that will sit there and not say a word, unlike my 19 year old daughter who says lots of words. 

So I cannot really answer your question. .  

[db16122]

ASM330LH seems to be better option since it has may application notes which means more application experience already.Good luck!

[Andrew Neil]

---

@DUTCHnotDUTCH wrote:

Apologies in advance for sounding like a noob (that is because I am). 

---

Welcome to the forum.

No need to apologise for being a "noob", but it would help to know what level of experience you do have.

Do you have any experience with microcontrollers? With programming in general? With electronics?

It's really clear exactly what  you need help with?

Interfacing multiple sensors is not essentially any different from interfacing one sensor - it's just a matter of distinguishing which is which.

So certainly start by getting just one sensor working before moving on to multiple.

What is reason for needing multiple sensors? Why do they need to be in a line?

The ASM330LHHTR and A3G4250DTR are both available as 24-pin DIL evaluation modules from ST:

https://www.st.com/en/evaluation-tools/steval-mki125v1.html

https://www.st.com/en/evaluation-tools/steval-mki193v1.html 

The Bluetooth is an entirely separate issue: it's just about transporting data - it neither knows nor cares what the data is, or where it comes from.

So, again, treat that separately.